Clang's qualifiers fail to parse in GCC, so __attribute__((nonnull (1)))
seems like the only somewhat portable solution.

Would you folks actually consider using it?

On Fri, Feb 24, 2023 at 12:07 PM enh <enh@google.com> wrote:

>
>
> On Fri, Feb 24, 2023 at 9:00 AM Jₑₙₛ Gustedt <jens.gustedt@inria.fr>
> wrote:
>
>> Tamir,
>>
>> on Fri, 24 Feb 2023 11:42:30 -0500 you (Tamir Duberstein
>> <tamird@google.com>) wrote:
>>
>> > Internal UB means "applying zero offset to null pointer" when the
>> > caller passes (NULL, _, 0, _).
>>
>> So you mean internal misbehavior?
>>
>> > Does such an attribute exist?
>>
>> Yes, gcc has the nonnull old-style attribute, and I think that the C23
>> versions of gcc and clang will have something like `[[gnu::nonnull]]`.
>>
>> Having this as attribute of the function and not of the pointer is a
>> bit unfortunate. For non-void pointers you could also use the `[static
>> 1]` array parameter notation instead of a pointer.
>>
>
> bionic's using clang's _Nullable and _Nonnull, which do let you express
> this:
> https://clang.llvm.org/docs/AttributeReference.html#nullability-attributes
>
>
>> Jₑₙₛ
>>
>> --
>> :: ICube :::::::::::::::::::::::::::::: deputy director ::
>> :: Université de Strasbourg :::::::::::::::::::::: ICPS ::
>> :: INRIA Nancy Grand Est :::::::::::::::::::::::: Camus ::
>> :: :::::::::::::::::::::::::::::::::::: ☎ +33 368854536
>> <+33%203%2068%2085%2045%2036> ::
>> :: https://icube-icps.unistra.fr/index.php/Jens_Gustedt ::
>>
>